Others have mentioned resonance as a cause for the AA587 and that was the cause in my estimation. It was not resonance of the vertical stabilizer that some here may have imagined causing the vertical stabilizer to break off, but instead, the entire aircraft oscillating in yaw around its vertical axis.
This oscillation reached an amplitude that you could never generate with a single application of full rudder. Instead it built upon the energy of preceding oscillations until the combined effect of all the oscillations plus increased angle of attack of the vertical stabilizer/rudder combination as the rudder reversed broke the vert stabilizer.
A resonant structural oscillation looks like this:
http://upload.wikimedia.org/wikipedi...-Resonance.PNG
Most of my flight time has been in what are now considered old jets with irreversible hydraulically powered controls that were connected to the stick and rudder pedals by cables and pushrods.
With irreversible controls, you do not have a clue to the forces you are exerting unless you have some sort of feedback. One rather fast aircraft that I flew used artificial feedback and changed its rudder pedal force from 2.6 lbs force/degree of travel to 11.5 lbs force/degree of travel at about 225 kt by means of an airspeed switch and a hydraulic centering cylinder. It then used the higher pedal force all the way up to its limit speed which exceeded M 2.0. The engineers were not at all concerned about us breaking the tail at that higher force ratio, but there were warnings to avoid excessive rudder deflection should the airspeed switch fail to work properly. Full travel of the rudder and rudder pedals was not restricted on that aircraft, but it was essentially impossible to exceed the intended rudder travel limits.
On the A306, the variable rudder travel limits protect the aircraft against a single rudder input overstress. Unfortunately, that system was not (and is still not) able to protect against an oscillatory rudder input where the input frequency approaches the natural tail wagging frequency of the aircraft. This system is also a mechanically controlled ,irreversible, hydraulic powered flight control with artificial feel. The problem is that the force and travel required to activate the rudder to its limits were minimal, and thus it was easy to excite a yaw oscillation by relatively small repetitive rudder inputs. Not the best design, but now that everyone is aware of the hazard, it is unlikely to bite again.
Incidentally, the simcrash posted by SMOC was almost certainly caused by resonance with the natural frequency of the simulator on its base.
http://virtualmystic.files.wordpress...6/simcrash.jpg
This could then have been prevented by a software change to attenuate near-resonant frequency inputs.